On testing for clusters using the sample covariance
β Scribed by Peter Bryant
- Publisher
- Elsevier Science
- Year
- 1975
- Tongue
- English
- Weight
- 517 KB
- Volume
- 5
- Category
- Article
- ISSN
- 0047-259X
No coin nor oath required. For personal study only.
π SIMILAR VOLUMES
A study of the consequences of using estimated prediction errors (residuals) instead of true (unavailable) innovations in covariance-based model validation tests reveals that a larger risk of underfitting than expected is obtained if residuals are used instead of innovations, and suggests means for
It is well known that the sample covariance is not an efficient estimator of the covariance of a bivariate normal vector. We extend this result to elliptical distributions and we propose a simple explicit estimator, which is efficient in the normal case and which outperforms the sample covariance in